Kaggle ITBench - SRE: leaderboard

Metric: Score (%). Source: www.kaggle.com. 13 models tracked.

Top models

#ModelScore
1Gemini 2.5 Flash50
2Gemini 3 Pro (Preview)50
3O350
4GPT-543.75
5GPT-5.143.75
6Claude Haiku 4.537.5
7Claude Opus 4.537.5
8Gemini 2.5 Pro31.25
9Grok 4.1 Fast (Reasoning)31.25
10Claude Sonnet 4.525
11Qwen 3 235B A22B 2507 Instruct18.75
12Mistral Small 3.112.5
13DeepSeek R1 05280

Interactive version: theaggregate.ai/benchmark?slug=kaggle-itbench-sre · How It Works · Data refreshed daily, snapshot 2026-09-19.